At a Glance:
Relaticle is a self-hosted, open-source CRM with a production-grade MCP server that connects AI agents like Claude or GPT to 30 CRM operation tools, a REST API, 22 customizable field types, and multi-team data isolation.
Overview:
Relaticle is a self-hosted CRM designed for developer-led teams, AI-forward startups, and SMBs that need AI agent integration with full data ownership. It pairs a complete CRM system with a Model Context Protocol server, allowing AI agents to perform CRM operations through 30 available tools. The data model supports 22 field types, including entity relationships, conditional visibility, and per-field encryption, without requiring migrations. A REST API exposes full CRUD operations with schema discovery for AI agents, while a five-layer authorization system enforces team-scoped data and workspaces. The interface is built on Filament 5 and Livewire 4, running on a Laravel 12 and PHP 8.4 stack. Relaticle ships in English but supports internationalization through language directory overrides.
Key Decision Points:
Self-hosted deployment: Runs on your own server via Docker or manual installation, with all data staying under your control.
MCP server built in: Includes a production-grade Model Context Protocol server with 30 tools, enabling Claude, GPT, and other MCP-compatible AI agents to interact directly with CRM data.
Team-scoped data isolation: A five-layer authorization model separates workspaces by team, ensuring each team only sees its own contacts, pipelines, and records.
No-migration custom fields: 22 field types – including entity relationships and conditionally visible fields – can be added dynamically without schema changes or deployment steps.
REST API with schema discovery: Exposes full CRUD endpoints, allowing external services and AI agents to discover available fields and interact with CRM records programmatically.
Core Features:
MCP server with 30 tools: Provides AI agents with standardized access to CRM operations, including data creation, updates, queries, and relationship management.
22 custom field types: Supports text, numbers, entity relationships, per-field encryption, and conditional visibility, all configurable without migrations.
Multi-team workspaces: Isolates data through a five-layer authorization system, scoping records, views, and pipelines to specific teams.
REST API: Offers full CRUD with schema discovery, enabling external applications and AI agents to understand available fields and interact with CRM records.
Per-field encryption: Allows individual fields to be encrypted at rest for sensitive data handling.
Use Cases:
Developer-led teams that want to give AI coding assistants or custom agents direct access to CRM data through an MCP server.
AI-forward startups needing a CRM where AI agents can create records, query pipelines, and manage relationships without third-party API gateways.
SMBs managing multiple teams that require strict data separation between workspaces while maintaining a single self-hosted instance.
Self-hosters who need a CRM with dynamic custom fields, entity relationships, and schema introspection available through both web UI and API.
Open-Source Alternative Value:
Relaticle is distributed under AGPL-3.0 and runs entirely on your own infrastructure, which means the source code is available for review, customization, and long-term maintenance without dependency on a vendor's hosting decisions. Its built-in MCP server lets AI agents operate on CRM data through a documented protocol rather than a proprietary integration layer. The dynamic field system avoids the schema lock-in common in SaaS CRMs, while internationalization overrides allow self-host forks to translate the application independently of upstream releases.